Benchmarking hospital marketing metrics that matter

Marketing gets roughly ten minutes in a hospital board meeting. In that window someone will show a cost per new patient that improved year over year, sitting next to a peer figure from a purchased benchmark set. Both numbers can be technically correct and still push the finance committee into approving a service line expansion that never pays back. This lesson settles three leadership questions: which of the metrics your team already produces earns a place on the board page, what good looks like once you compare across systems and countries, and how the comparison gets gamed before it ever reaches you.

The four numbers that reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition → the board

A board page carrying eight metrics carries none. Four, each with a named owner and a trailing four-quarter trend, is the working limit. Everything else stays at operating level.

1. Cost per new patient, by service line, on completed visits

Assume the per-service-line calculation the acquisition lesson builds. What leadership arbitrates is the reporting standard wrapped around it, because that is where the number actually moves:

Board-reported cost per new patient =
    all marketing cost (media + agency + in-house salaries + marketplace fees)
    / patients with a completed first visit
      (36-month lookback, net of the baseline a holdout or geo test implies)

Three choices hide in there. A 12-month lookback instead of 36 reclassifies returning patients as new and flatters the figure without a single additional person walking through the door. The cost perimeter is the second: a system with a 20-person in-house studio can report media-only spend and look half as expensive as a peer that outsources the same work to an agency. The third is the one that costs real capital. Brand-keyword and retargetingretargetingShowing ads to users who have previously visited your site or interacted with your brand, to bring them back and drive conversion.View full definition → spend captures people who had already decided; systems that run paused-market or geographic holdout tests commonly find their incremental cost per patient is a multiple of the blended figure they had been reporting. If your board number has never been holdout-tested, say so on the slide.

2. Value-to-cost ratio, by payer or contract type

Assume the multi-year contribution model the lifetime valuelifetime valueLifetime Value: the total revenue (or profit) a customer generates throughout their entire relationship with your business.View full definition → lesson builds. The 3:1 floor that circulates in hospital marketing decks was imported from subscription software, and it fails against two hospital realities.

Capacity is the first. If the orthopaedic theatre runs near full utilisation, the marginal acquired patient displaces one who would have arrived anyway, and the ratio measures queue position rather than growth. Contract type is the second. Intermountain Health operates its own health plan, SelectHealth, and carries risk on attributed lives; for that population the arithmetic inverts, since a newly acquired high-utilisation patient adds cost. The number that belongs on Intermountain's page is attributed lives and care-gap closure, not acquisition volume. In Germany, Helios Kliniken (part of Fresenius) works inside case volumes negotiated with sickness funds, and volume above the agreed level is reimbursed at a discount, so share gains beyond budget can be margin-negative. One campaign, three different verdicts.

3. One funnelfunnelThe customer journey from awareness to purchase, typically Awareness, Interest, Consideration, Decision, Action, with prospects narrowing at each stage.View full definition → step, named, with an owner

The funnel lesson teaches how to find the leaking step. The board sees exactly one: the worst step, the name of the person fixing it, and a date. Boards that review six conversion rates end up managing the call centre from the boardroom, which is the fastest way to lose the operator's ownership of it.

4. Continuity, set against new patients

Assume the continuity and lapse measures the retention lesson owns. On this page they exist to stop acquisition growth being celebrated while the base drains. A system adding 4% new patients a year while losing 6% of its continuing relationships is shrinking, and only the two numbers side by side make that visible.

What good looks like across systems and geographies

Mayo Clinic's consumer health site draws tens of millions of visits a month, most from people searching symptoms with no intention of travelling to Rochester. Its visit-to-enquiry rate is structurally low, and any hospital that copies Mayo's funnel ratios is comparing a publishing operation with a demand-generation one. The same brand strengthbrand strengthThe commercial value your brand adds beyond functional product attributes: the price premium, preference and loyalty it generates.View full definition → bites from the other direction: destination referrals arrive with high case value and consideration periods measured in months, so a 30-day attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → window credits marketing with almost nothing it earned.

IHH Healthcare runs Gleneagles and Mount Elizabeth in Singapore, Acibadem in Turkey and Fortis in India, and a material share of revenue at its Singapore and Turkish hospitals comes from international patients. Cost per acquired patient there behaves like travel marketing: agent commissions, visa and translation support, long decision cycles, case values well above the domestic equivalent. Track Acibadem's cost per patient in lira across years without adjusting for Turkish inflation and the trend is fiction. Cross-border comparison inside one group needs a common currency, a common definition of a new patient, and a note on what each market's payer structure allows before the figures sit in the same column.

Advertising rules also move the ceiling. German promotion of medical treatment operates under statutory advertising restrictions and physician professional codes, which narrows Helios's channel mix compared with a US system bidding freely on procedure keywords. A German cost per patient that looks admirably low may reflect regulation, not skill.

Benchmark against a cohort you can name in one sentence: same competitive density tier, same funding structure, comparable case mix. If you cannot list the four systems in your cohort, you have a national average, not a benchmark. Your own trailing history stays the primary comparison; external figures are the sanity check. The Content Marketing Institute publishes free, non-vendor guidance on measurement discipline that you can adapt.

How the comparison gets gamed

• Definitions drift mid-year. The lookback shortens, the denominator switches from completed visits to bookings, and the improvement is entirely clerical.
• Service lines get reclassified. Move joint-injection patients into orthopaedics and the line's cost per patient falls while nothing clinical changes.
• Geography gets cherry-picked. One high-performing metro is presented as the system result.
• Reactivated patients are counted as new, which double-charges the base for growth it already provided.
• The reported funnel step rotates: whichever step improved this quarter is the one shown.

Putting it together: the board page

Read it as an allocation decision, not a performance review. A line whose cost per patient looks strong but has never been holdout-tested gets a test, not more budget. A line at capacity gets operational investment before media. A line where continuity erosion outruns acquisition gets its spend frozen until the base stabilises.

Common benchmarking mistakes

• Buying a benchmark set and skipping the cohort definition, then treating the gap as a performance problem.
• Tying incentive compensation to a ratio nobody has audited. Bonus-linked cost per patient reliably drives spend toward the cheapest conversions, which are usually patients who were already coming.
• Comparing across borders without normalising funding structure, advertising law and inflation.
• Presenting a system-wide average, which hides both the line that should double its spend and the line that should stop.
• Letting the marketing team own the definition. The finance function should sign the denominator once a year and keep it fixed.

Key takeaways

• Four numbers, four owners, one page. The board arbitrates allocation; the operating team keeps the diagnostic detail.
• A cost per new patient that has never been holdout-tested is a spend record, not a performance measure.
• Value-to-cost floors are contract-dependent. Under capitation or negotiated volume caps, acquisition growth can destroy margin.
• Cross-border and cross-brand comparisons need a common currency, a common new-patient definition and a note on advertising rules before they belong in one table.
• Most benchmark improvements you will be shown are definitional. Freeze the definitions annually and restate history when you change them.